aacenc: avoid double in quantize_bands.
authorReimar Döffinger <Reimar.Doeffinger@gmx.de>
Tue, 1 Mar 2016 21:53:18 +0000 (22:53 +0100)
committerReimar Döffinger <Reimar.Doeffinger@gmx.de>
Sun, 6 Mar 2016 10:49:19 +0000 (11:49 +0100)
commit0a04c2885f02f7db6b410b6d43d120e5eb48dc18
tree9a093267f48e45552489606d62bd2518e7598ef4
parent7dd60a0b54ee2d4a5aa2a069597b560e82afb201
aacenc: avoid double in quantize_bands.

I cannot see any point whatsoever to use
double here instead of float, the results
are likely identical in all cases..
Using float allows for much more
efficient use of SIMD.

Signed-off-by: Reimar Döffinger <Reimar.Doeffinger@gmx.de>
libavcodec/aacenc_utils.h